PROJECT SUMMARY/ABSTRACT The hippocampal formation is critical for the formation and retrieval of autobiographical memories, which consist of information about what happened when and where. It is known how objects, context and space (what, where) are represented in the hippocampus and it has recently been described that time (when events occur) on a scale of seconds to minutes is represented by the sequential activation of hippocampal cells. On a longer time scale, a time-varying neural code has previously been shown by theoretical models to be suitable for estimating the recency of a remembered event. In recently published results and results presented as preliminary data, we identify a novel, time-varying hippocampal neural code that can represent how long ago an event occurred on a time scale of hours and days. Our data first identified that the neuronal firing patterns of CA1 cells are characterized by a monotonic accumulation of rate differences as a function of time between experiences. However, we demonstrate that stored information does not simply deteriorate in the CA1 area, but that the code for time can co-exist with reliable coding for other aspects of an event, such as the spatial location or the context. We also found that CA3 contains an exquisitely precise code for context and space that does not vary over time. New preliminary data show an effect in CA2 that is opposite of CA3. CA2 cells represent elapsed time but no information about context. Based on these findings, we hypothesize that the neuronal code for extended time is combined with spatial and contextual information in the CA1 cell population to guide behavior in which the recency of a previous event is remembered. Using a combination of behavioral testing and single-unit recordings in awake behaving rodents, this hypothesis is tested in three aims that: (1) determine whether the neuronal code for extended time intervals is generated in the hippocampal CA2 neural network, (2) determine whether the input from the hippocampal CA3 subregion is necessary for maintaining the stable memory coding component in the CA1 neural network over hours and days, and (3) determine whether a time-varying neural code in hippocampal CA1 and CA2 subregions is correlated with remembering how long ago. The first two aims address how different subregions contribute to the time-varying and stable components of the neural code. In addition to defining the function of CA2 for temporal coding and the function of CA3 in enabling stable memory representations, we will also ask whether the entorhinal cortex can represent time on an extended scale and is thus a brain region in which temporal coding over a large range of scales can be found. In the third aim, we will use a behavioral task in which it has been shown that rats remember how long ago over extended time periods. We will measure the similarity of activity patterns in place cell populations between two time points and determine whether the change in neuronal firing patterns corresponds to the rat's estimate of elapsed time. Taken together, these studies will be important for understanding the neural network mechanisms for long-term memory stability and temporal event coding in the brain structures that support episodic memory. Understanding the key mechanisms for memory processing will guide efforts to repair circuit dysfunction in psychiatric, neurological, and neurodegenerative diseases.